Skip to main content
Donate to the Python Software Foundation or Purchase a PyCharm License to Benefit the PSF! Donate Now

submit jobs to LSF with python

Project description


python wrapper to submit jobs to bsub (and later qsub)

@brentp, @brwnj


>>> from bsub import bsub
>>> sub = bsub("some_job", R="rusage[mem=1]", verbose=True)

# submit a job via call'ing the sub object with the command to run.
# the return value is the numeric job id.
>>> print sub("date").job_id.isdigit()

# 2nd argument can be a shell script, in which case
# the call() is empty.
#>>> bsub("somejob", "", verbose=True)()

# dependencies:
>>> job_id = bsub("sleeper", verbose=True)("sleep 2").job_id
>>> bsub.poll(job_id)



For file jobs, we can emulate shell syntax:


job = bsub('my-job') < ''

Same for text commands:


"echo hello" | bsub('other-job')



It's possible to specify dependencies to LSF using a flag like:

bsub -w 'done("other-name")' < myjob

We make this more pythonic with:


>>> j = sub('sleep 1').then('sleep 2')

which will wait for the first job `sleep 1` to complete
before running the second job `sleep 2`. These can be chained as:


j = sub('myjob')
j2 = j('sleep 1')
j3 = j2.then('echo "hello"')
j4 = j3.then('echo "world"')
j5 = j4.then('my scripts.p')

# or:

j('sleep 1').then('echo "hello"').then('echo "world"')

Where each job in `.then()` is not run until the preceding job
is `done()` according to LSF.

Bioinformatics example of chaining:

This would submit jobs for positive and negative strand coverage in parallel.
Each strand submitting jobs that run serially.


from bsub import bsub

submit = bsub("bam2bg", verbose=verbose)

# convert bam to stranded bg then bw
sample = "subject_1"
chrom_sizes = "chrom_sizes.txt"

# submit jobs by strand for parallel processing
for symbol, strand in zip(["+", "-"], ["pos", "neg"]):

bigwig = "" % (sample, strand)
bedgraph = "%s_%s.bedgraph" % (sample, strand)

bam_to_bg = ("bedtools genomecov -strand %s -bg "
"-ibam %s | bedtools sort -i - > %s") % (symbol, bam, bedgraph)
bg_to_bw = "bedGraphToBigWig %s %s %s" % (bedgraph, chrom_sizes, bigwig)
gzip_bg = "gzip -f %s" % bedgraph

# process strand-based steps serially
# submit first 2 jobs to default queue; final job to 'gzip' queue
submit(bam_to_bg).then(bg_to_bw, job_name="bg2bw").then(gzip_bg, "gzipbg", q='gzip')



use the command-line to run jobs with auto-specified err and log files:

echo "hello" | python -m bsub -J "fake"
bsub -J fake -e fake.%J.err -o fake.%J.out < /tmp/
If a log/ directory exists, the logs will be placed there.

the shell script is automatically created and cleaned up after use.

Project details

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Filename, size & hash SHA256 hash help File type Python version Upload date
bsub-0.3.5.tar.gz (7.0 kB) Copy SHA256 hash SHA256 Source None

Supported by

Elastic Elastic Search Pingdom Pingdom Monitoring Google Google BigQuery Sentry Sentry Error logging AWS AWS Cloud computing DataDog DataDog Monitoring Fastly Fastly CDN SignalFx SignalFx Supporter DigiCert DigiCert EV certificate StatusPage StatusPage Status page